A cross-sectional design was used to collect survey data in a web-based questionnaire. The effects of age, gender, and region of residence on severity of symptoms were further explored. The 15-min survey included additional questionnaires; however, those characteristics were outside the scope of the present analysis. The data was collected between April 8 and June 1 2020. The timeline of this study within the context of COVID-19 development is depicted in Figure 1.

Participants were recruited via flyers, social online platforms, mailing lists and notices in in-patient clinics and supermarkets. Out of 2,506 participants who had started the questionnaire, 1,739 datasets (69.4% valid sets; 72.6% female) contained data that was usable for statistical analysis for this manuscript. We excluded 280 participants who reported to have been or were in psychotherapy, as the sample we used for comparison (reference sample) was a mentally healthy group of individuals. The resulting sample size was N = 1,459 (71.4% females), the age is ranged between 18 and 88 years, M = 34.35, SD = 14.04. Exclusion criteria were an age younger than 18 years, unrealistic or missing values as well as not having completed the sections needed for this analysis. Demographic data are depicted in Table 1. An indication of the e-mail address for future investigations was voluntary, otherwise no further personally identifiable information was collected. All participants gave their informed consent for participation and completed the questionnaires electronically. Data was collected anonymously without IP addresses or GPS tracking. Email address, when provided, was separated from the rest of the dataset. The study was approved by the Ethics Committee of the Department of Psychology at the PFH – Private University of Applied Sciences (Ethics application number: 251982).

The subsequent analysis focused on comparing our sample and the reference sample. When comparing means between independent groups we calculated Welch’s t-tests, due to inhomogeneity of variance between samples. As a measure for effect size, we used Hedge’s g due to a large difference in sample size. Furthermore, for depressive symptoms, we explored differences between male and female participants, age groups and states. In a series of hierarchical regression models, combinations of age, gender and state were used to predict the average score of depression, compulsion, anxiety, eating disorder and somatization symptoms. To correct for multiple comparisons where appropriate, the false discovery rate [FDR, Benjamini and Hochberg (1995)] was applied. Reported p-values reflect the FDR correction. The statistical analysis was carried out using R 3.6.2 (R Core Team, 2017).

The items with the highest mean scores in the survey sample were “I feel down and depressed” (M = 1.25, SD = 1.04) for the depression subscale, “I try to avoid these harmless frightening situations” (M = 0.76, SD = 1.09) for the anxiety subscale, “I try to resist recurring, seemingly senseless thoughts and actions, but often don’t succeed” (M = 0.54, SD = 1.00) for the compulsion subscale, “I worry about having a serious physical illness” (M = 0.37, SD = 0.80) for the somatization subscale and “I spend a lot of time thinking of ways to lose weight” (M = 0.90, SD = 1.21) for the eating disorder subscale.

We found significant differences between the reference and the COVID-19 sample in depressive symptoms, compulsive symptoms and symptoms of eating disorder, as displayed in Table 2 and Figure 3, with higher burden in the COVID-19 sample. We found no different scores on the anxiety scale and a very small difference in somatization symptoms with higher prevalence in the reference sample (see Table 2). Significant medium effects for depressive disorder and significant but small effects for compulsive, eating disorder and somatization symptoms were found.

We calculated linear regression models for each of the ISR symptom scales (N = 1,459; see Table 3 and Figure 4). For each subscale, we compared models that only considered main effects to models that included an interaction term. Except for the eating disorder subscale, we found that the added interaction term was neither a significant predictor nor did it improve the model fits. In all subscales except somatization, age was a significant predictor. For instance, for every additional year in participants’ age, the depression score decreased by 0.012 points. Similarly, male participants reported less severe symptoms than female participants (again, except for somatization). We then correlated the subscales against each other (Figure 5). Except for the eating disorder subscale, all subscales were significantly correlated to at least one other subscale, with the compulsion and anxiety subscale showing the strongest correlation.

There were no significant differences in average symptom severity for depression, compulsion, anxiety, somatization or eating disorder symptoms between Lower Saxony and Bavaria (all p > 0.155). Average male scores were at M = 0.79 (SD = 0.72) in Lower Saxony and M = 0.87 (SD = 0.80) in Bavaria, average female scores were M = 1.01 (SD = 0.90) in Lower Saxony and M = 0.90 (SD = 0.78) in Bavaria. For eating disorders, average scores were at M = 0.75 (SD = 1.00) in Lower Saxony and M = 0.76 (SD = 1.03) in Bavaria. Average male scores were at M = 0.46 (SD = 0.70) in Lower Saxony and M = 0.53 (SD = 0.86) in Bavaria, average female scores were M = 0.88 (SD = 1.08) in Lower Saxony and M = 0.85 (SD = 1.08) in Bavaria. Means and standard deviations can be found in Table 4.

A general increase in the rates of depressive symptoms during a pandemic would be in line with previous studies conducted before the coronavirus outbreak (Hawryluck et al., 2004; Yip et al., 2010; Brooks et al., 2020) and with other prevalence studies conducted during the COVID-19 pandemic (Bäuerle et al., 2020b; Pieh et al., 2020; Rossi et al., 2020; Wang et al., 2021) as well as longitudinal studies (Castellini et al., 2021). With regard to our exploratory findings, starting with eating disorder symptoms, two studies investigating reactions of formerly diagnosed patients also reported worsening of eating disorder symptoms during lockdown in European samples (Fernández-Aranda et al., 2020; Robertson et al., 2021), studies targeting eating disorder symptoms in the general population during the COVID-19 crisis are still scarce to our knowledge: One Australian study showed an increase in restricting and binge eating behaviors (Phillipou et al., 2020). Furthermore, increases in unhealthy eating behavior became apparent: One study conducted in Italy found a weight gain in 48.6% of the population (Di Renzo et al., 2020) and an international online survey found decreases in physical activity and more unhealthy food consumption patterns (Ammar et al., 2020). There are not many studies with a focus on compulsive symptoms, however, in former patients (Jelinek et al., 2021) as well as adolescents and children (Tanir et al., 2020), increases in symptomatology have been documented. Interestingly, our study did not find elevated anxiety symptoms, which is contradictory to the results of, e.g., Li et al. (2020) and Zhu et al. (2020). However, other studies also showed no increase in anxiety symptoms in China (e.g., Wang et al., 2020) and in a longitudinal study in the Netherlands (Pan et al., 2021). Higher anxiety levels have been shown to be related to lower social capital (Xiao et al., 2020), to poor mental health (Wang et al., 2021) and to COVID-19-infections (Özdin and Bayrak Özdin, 2020) while living in urban areas and a steady family income served as protection from elevated anxiety levels (Cao et al., 2020). As most people in Germany have comparably high financial security, the degree of COVID-19-infections in our sample was low and only mentally healthy participants were included in this study, our sample might therefore not have been very vulnerable to experience anxiety. Most surprising was the low prevalence of somatization symptoms. When looking at the intercorrelations of the scales, it was interesting to note that this study showed low associations between the depressive subscale and the other scales, where other studies show high associations (e.g., Olfson et al., 2017). However, one study recently demonstrated that there are specific risk factors, not only predicting elevated psychopathological symptoms during the COVID-19 pandemic, but also comorbidities (Palgi et al., 2020). This study showed that loneliness predicted the comorbidity between anxiety and depressive symptoms. Furthermore, one could assume, that the increase in depressive symptoms was a normal rather than a psychopathological reaction to a loss of incentives, social support, ease of life, and very specific to the restrictions following the COVID-19 pandemic and therefore not associated with other psychopathologies.

Regarding gender and age, some differences in the rates and severity of psychological burden were found: Females showed higher rates of all scales except for somatization, which was not surprising considering that the general risk of those disorders is higher for this group (Sepulveda et al., 2008; Busch et al., 2013; Grobe et al., 2018). Females might also be especially vulnerable during the COVID-19 pandemic because they often tend to work in fields that were most affected by the pandemic, they are also more likely to be responsible for childcare which might increase worrying behaviors (Alon et al., 2020; Niedzwiedz et al., 2021). Furthermore, lockdown increased risks of domestic violence, which usually affects women more often (Alon et al., 2020).

Our results suggest, that younger people have higher symptoms in all symptom subscales than older participants, except for the somatization scale. This supports previous findings (Bäuerle et al., 2020b; Pieh et al., 2020) that mental health of young adults and women was significantly more burdened by the COVID-19 pandemic and the lockdown. One explanation could be that older people often have better emotion regulation skills (Charles, 2010) that help coping with the psychological effects of the pandemic (Barber and Kim, 2021). Results from this study showed that older men exhibited less worrying related to the Corona pandemic. Pieh et al. (2020) assumed that younger people are usually more likely to experience job insecurity and financial problems, the lockdown and the resulting restrictions might therefore have a greater impact on the daily lives of younger adults. It is important to note, that, although significant, these were small effects and our model could only explain 4% of variance.

Contrary to other studies (e.g., Benke et al., 2020), no difference in symptomatology was found between people living in Lower Saxony and Bavaria even though Bavaria had a stay-at-home order and Lower Saxony did not. This result, however, is consistent with data collected conducted prior to the pandemic, where no significant difference in prevalence rates for depression in adults was found between these two states regardless of gender (Bretschneider et al., 2017). Therefore, it can be concluded that a voluntary stay-at-home order in comparison with a forced stay-at-home order did not have an impact on differences in mental health symptoms in our sample.

There are several limitations to this study. First, any observations made here are purely correlational. There were neither baseline measures nor any follow-up measures conducted and therefore no statements about the development of symptoms or long-term effects can yet be made. Second, the data presented here was collected from a convenience sample. Therefore, even despite our resampling efforts, the sample reported on here, likely differs from the reference sample (Tritt et al., 2010). As the study was mainly conducted with residents of Lower Saxony and Bavaria, it is unclear if the data can be generalized to Germany’s population as a whole. In addition, the reference sample used to establish differences in symptom rates, is ten years old (Tritt et al., 2010). Since then, an increase in depressive disorders is likely, as incidence rates for depressive disorders in Germany have increased (see Kaufmännische Krankenkasse., 2020, May 25). Our study therefore cannot clearly state whether the high symptom prevalence is due to the restrictions or due to a general increase in symptoms during the last decade. Second, despite the described differences in lockdown measures between the two German states, in both states restrictions were imposed. Therefore, the study can mainly make the specific point that a forced stay-at-home order did not have an impact on psychological burden compared to a voluntary stay-at-home order.

Still, the results of this study complement the results of other research during the first weeks after restrictions: It matches the results of comparable studies in other countries in terms of depressive symptoms (e.g., Pieh et al., 2020) and provides interesting new results concerning high prevalence of eating disorder and compulsive symptoms, especially for younger and female adults. As the ISR has been validated with a large sample size and shows good quality criteria, it serves as a good measure of psychological burden, so the data here can be considered as a screening. The current study could therefore be interpreted as a warning sign. The duration and causes of psychological burden should be further investigated, and countermeasures should be taken. In a nationwide lockdown, online mental health counseling is a far-reaching method to help many people (Dan, 2020). One helpful prevention measure is to educate the public on what mental health issues they may face during times of isolation. It is useful to provide tips on how people can handle individual situations and suffering, such as loneliness, in order to manage these situations and minimize suffering (Hiremath et al., 2020). Tele-mental-health services, online platforms and social media are helpful for support and care during mental health crises in the Corona pandemic, especially in rural areas (Zhou et al., 2020). For instance, the e-mental health intervention ^‵CoPE It^‵ is a low-threshold approach to support people with mental distress during the Corona pandemic (Bäuerle et al., 2020a). In Wuhan, a free psychological counseling service was offered online to reduce people’s distress (Dan, 2020). Employers in Canada supported their employees by dropping mental health fees, providing support and educating them about mental health (Ho et al., 2020). Furthermore, various online mental health services can be taken as preventive measures to identify at-risk groups (Liu et al., 2020). Still, for those developing mental disorders face-to-face therapies should be available. Sports (Pieh et al., 2020) as well as pursuing new projects or hobbies at home (de Quervain et al., 2020) seemed to function as resilience factors.